Ann Marie Holtrop

Modeling changes in freshwater mussel diversity in an agriculturally dominated landscape

Abstract.  Freshwater mussels perform critical ecosystem functions and provide many valuable ecological services. However, anthropogenic effects have severely decreased mussel diversity and abundance at both local and regional scales. Understanding how human disturbances, particularly landuse change, and fish assemblages are related to mussel assemblages is essential for effective conservation and restoration. We used Random Forests (RF) regressions, a data-mining technique, to examine how mussel species richness, total abundance, and abundances of individual species were related to land use at different spatial scales and to fish species richness and abundance in east central Illinois, USA. Mussel richness increased with % wetlands, % open water, % grassland in the riparian zone, and total fish abundance; decreased with % urban land in the riparian zone; but responded poorly to fish species richness and fish Index of Biotic Integrity (IBI) scores. Total mussel abundance mainly increased with total fish abundance and decreased with both % urban land in the riparian zone and road density. Of 8 mussel species modeled, the abundances of 3 were strongly related to total fish abundance or fish-host abundance, 3 with both fish abundance and land use, and 2 with land use and other physical variables. These findings can help researchers and resource managers explain the spatial variation of mussel assemblages and choose abiotic and biotic variables to monitor or manipulate for maintaining or restoring overall mussel diversity or the populations of individual species.

Reconstructing the natural distribution of individual unionid mussel species and species diversity in wadeable streams of Illinois, USA, with reference to stream bioassessment

Freshwater mussels are considered one of the most imperiled aquatic species groups in the USA. One of the challenges in protecting and restoring mussel biodiversity effectively is a lack of understanding of their natural distributions at relevant spatial scales. Without this information, assessing the current status of individual species and overall biodiversity or evaluating restoration success is difficult. We compiled records for 45 mussel species and a range of natural environmental variables describing climate, geology, soil, land cover, and watershed topography in Illinois wadeable streams. We used reaches (segments between 2 neighboring tributaries) as the basic spatial unit of the stream network for modeling species distributions with Maxent. We applied these models statewide to all identified wadeable reaches. Stacking the predictions of individual models yielded an estimate of species richness for each reach. The estimates were compared with observed richness from 2 sets of independent sites: 17 sites sampled multiple times over the past 50 to 100 y and 18 sites intensively sampled in 2009 to 2010. The latter set was expected to represent much more impaired mussel assemblages than the former. These sets of sites lost an average of 25 and 46%, respectively, of species expected under natural conditions. Observed occupancy of individual species decreased by 27 and 35%, respectively, from expected natural values. Listed species suffered heavy occupancy loss more frequently than nonlisted species. Estimated mussel species loss was negatively correlated with 2 existing indices of biological integrity. These results, together with maps of natural distributions for individual species and for species richness, will assist with mussel biodiversity conservation and the development and use of mussel-based biological indicators in stream assessment.